[R] Use of substitute -- was More powerful than objects() or ls()

[Prof Brian D Ripley]

On Tue, 27 Mar 2001, Patrick Connolly wrote:

> According to Prof Brian D Ripley:
> |>
> |> On Mon, 26 Mar 2001, Patrick Connolly wrote:
> |>
> .....
>
> |> > I think it would be simple enough if I could get something in R to
> |> > work the way substitute() does in Splus.  The thinking seems to be
> |> > very different between the two dialects.  Maybe if I tinker a bit
> |> > longer, I'll get the connexion between them.
> |>
> |> They are almost the same. (See V&R's `S Programming' for a detailed
> |> comparison.)   There is a difference when using function arguments:
> |> S uses the initial values, R the current ones, so in R one often needs to
> |> substitute early or take care to work on copies.
>
> Indeed, V&R explains it, so that's useful.  Thanks again, Brian.  But
> as Thomas Lumley anticipated, it is not straighforward to make use of
> comment with it.  Writing to and sourcing from a temporary text file
> is not very elegant, but also not very complicated,

Using output text connections would be more elegant (implemented in R only
at present, and only at 1.2.0 so post V&R).

> |> Ah, I think most of us keep our functions in files not in objects, even in
> |> S-PLUS. I even keep my datasets in files (or databases).  I know John
> |> Chambers advocates an `S object is master' view in the Green book, but I
> |> don't find it very practical given current tools.
>
> It is very practical with my Splus function (even though I could write
> it rather a lot better now).  It's useful on those occasions when I
> attach to a directory where I did similar work some years ago and look
> through the lists, functions, matrices, dataframes and the like that I
> used there. I can quickly decide what I can use in the present
> directory.  Even so, there is good reason to have a file of the code
> the function uses, and now I know how to use substitute, that's easy
> to automate.
>
>
> |> A CVS archive of R/S code is a very useful way of version control.
>
> Sounds interesting.  Where will I find information about CVS archives?
> V&R don't have any mention in the index/es.

www.cvshome.org, I think currently (was www.cyclic.com).  It's not to do
with R/S but a way to keep on a server files and records of all changes.
For example the R source code is kept in a CVS archive, as is all the V&R
material (tex files as well as software).

I guess the R-recommended way of working is to use packages, kept in source
form, and to archive the package source.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._